The aberrant assembly of normally soluble proteins into ordered aggregates, called amyloid fibrils, is a cause or associated symptom of many different human disorders including Alzheimer's, Parkinson's and Huntington's diseases, the prion diseases and adult-onset diabetes. Known collectively as the amyloidoses, these diseases are characterized by the slow deposition of a specific protein into amyloid fibrils, which then accumulate into plaques, destroying the function of the affected tissue, usually with degenerative and ultimately fatal consequences. An understanding of the molecular-level mechanisms that result in the aggregation of proteins into amyloid is essential for the discovery of potential therapeutic strategies and diagnostics. As-part of our NIH-sponsored effort to uncover the general physical principles that govern protein aggregation, we developed an intermediate-resolution protein model that allowed the simulation using discontinuous molecular dynamics (DMD) of multi-protein systems within days on a fast workstation. A recent breakthrough enabled us to simulate assembly of 48 16-residue alanine peptides into a fibrillar structure starting from the random coil state. This suggests that the intermediate-resolution model could be used as a computational tool to explore fibril formation in short peptides. The project has three specific aims: (1) to learn the basic physical principles governing protein fibril formation by using DMD to simulate multi-protein systems containing polyalanine chains modeled using our intermediate-resolution model, (2) to shed light on the molecular-level mechanisms responsible for the aggregation of polyglutamine, the protein whose fibrillization is linked to Huntington's disease, by extending the intermediate-resolution model to the treatment of polyglutamine side chains and then performing DMD simulations, and (3) to investigate the aggregation and possible fibrillization of multi-protein systems containing specific amyloidogenic peptides, particularly the Alzheimer's peptides Abeta(1-40) and Abeta(1-42), by extending the intermediate-resolution model to a coarse-grained representation of all 20 residues, performing DMD simulations, and comparing our results with experiment. This work should culminate in a detailed molecular-level picture of the fibrillization process, thereby providing insights to guide medical research workers directly involved in developing therapeutic strategies or inhibitors to circumvent those steps in the fibrillization process that are most responsible for cell damage.